A major challenge for current biology is to understand how bio-molecules interact and cooperate to achieve their joint cellular functions. Understanding a pathway requires that we first reconstruct the network - identify what molecules participate in a given pathway and how they interact, and then characterize its functional modes. Following the explosion of genome wide assays, network reconstruction became a central task in molecular biology, spanning signaling, metabolic, and regulatory pathways. However, despite intensive research, many fundamental cellular pathways are only partially known, and not well understood. For instance, in the metabolic pathways of S. cerevisiae, perhaps the best understood biochemical network, 20 percent of enzymes and transporters are unidentified. In the same organism, only 30 percent of the interactions in the signaling pathways are known. The goal of the research proposed here is to develop computational and statistical methods for completing partially known biological pathways and use these methods to reconstruct and understand two important molecular pathways: metabolic pathways and signal transduction pathways. This will be achieved by learning how diverse functional data maps onto known networks, identifying recurring patterns of this mapping, and use them to predict unknown network components. We have recently showed how these patterns, which we call “activity motifs”, can be identified from data, and reveal organization principles of metabolic and signaling pathways1-2.
Call for proposal
See other projects for this call